JP5799219B2 - Object detection device - Google Patents

Description

  The present invention relates to an object detection device that detects an object by obtaining a difference between a current thermal image and a background image using a thermal image.

  2. Description of the Related Art Conventionally, there has been provided an object detection device that acquires a thermal image of a detection range from a sensor device and detects a heat source such as a person as a target from the thermal image. As a method for detecting an object by the object detection device, a thermal image of the detection range at the time when no object is present in the detection range is recorded in advance as a background image, and the current thermal image and background image acquired from the sensor device are recorded. A background subtraction method for detecting a difference between and is known. The background subtraction method is advantageous in that the calculation is simple and the detection accuracy is good in a stable environment.

  By the way, if an object detection device using the background subtraction method continues to use a background image that has been acquired once, it will be erroneous when the temperature within the detection range that does not depend on the object (hereinafter referred to as “environment temperature”) changes. Since there is a risk of detection, it is necessary to update the background image.

  Therefore, an apparatus having a function of updating a reference image (background image) with an original image (acquired image) when a detected object is not detected has been proposed (for example, see Patent Document 1). Also, a function that captures an input image at a time interval that is a predetermined time interval from the time when a previous reference image (background image) is captured, obtains an average value of the input image and the reference image, and updates it as the next reference image An apparatus having the above has also been proposed (see, for example, Patent Document 2). However, in the apparatus described in Patent Document 2, when an object is detected from an input image, the addition averaging process is not executed.

  Furthermore, as an apparatus for identifying a human body from an image, an apparatus having a function of performing a difference operation between a background image and a current image, binarizing with a predetermined threshold, and rewriting the background image with the current image for pixels below the threshold Has also been proposed (see, for example, Patent Document 3). The device described in Patent Document 3 also has a function of rewriting a background image with a current image for a pixel that is determined to be not a human body even if it is a pixel that is equal to or greater than a threshold value (that is, a pixel detected by binarization). It is equipped.

JP-A-62-240823 Japanese Patent Laid-Open No. 10-308939 Japanese Patent Laid-Open No. 5-118916

  However, in the configurations described in Patent Documents 1 and 2, since the background image is not updated in the image in which the target object is detected, the background image is updated while the target object continues to exist in the detection range. Otherwise, it may not be possible to cope with a change in the environmental temperature, and erroneous detection may occur.

  Further, as in the apparatus described in Patent Document 3, even in the case where the background image is updated for an area in which the object (person) is not detected in the image in which the object is detected, the area in which the object is detected is Since the background image is not updated, it is not possible to cope with changes in the environmental temperature, and erroneous detection may occur. That is, for example, if the environmental temperature drops while a person continues to be at the same position, the human body temperature also decreases according to the environmental temperature, but the background image of the area where the person is present is updated while the environmental temperature is high. Not. In this case, in any case, the difference value between the human body temperature and the background image falls below the threshold value, and the object detection device may be erroneously detected (not reported) when there is no person. On the other hand, if the ambient temperature rises while a person is still present at the same position, the background image of the area where the person is present remains lower than the actual ambient temperature, and even if the person leaves that area, the object detection device May falsely detect that there are people in the area.

  The present invention has been made in view of the above-described reasons, and an object thereof is to provide an object detection apparatus that is less likely to cause erroneous detection by responding to changes in environmental temperature.

  The object detection apparatus according to the present invention stores an acquisition unit that acquires acquisition information representing a temperature distribution within a detection range, and background information that represents a temperature distribution within the detection range when no object is present in the detection range. A storage unit; a detection unit that detects the presence or absence of the object in the detection range based on a change in the acquired information with respect to the background information; and an update unit that repeatedly updates the background information stored in the storage unit; The update unit divides the background information into a detection region including a region where the object is detected by the detection unit and a non-detection region including a region other than the detection region, and the non-detection region For the first background update process for updating the background information based on the acquired information, and for the detection area, the first of the extraction area consisting of at least a part of the non-detection area Using the correction value obtained from the amount of change in temperature by Jing update process, and executes a second background update processing of updating the background information.

  In this object detection apparatus, the update unit obtains the correction value from the amount of change in temperature of the extraction region, with the partial region in contact with the detection region as the extraction region in the non-detection region. desirable.

  In this object detection apparatus, the update unit uses a partial area of the non-detection area of the background information as the extraction area, and obtains the correction value from the amount of change in temperature of the extraction area; The second mode for obtaining the correction value from the amount of change in the temperature of the extraction region, using the entire non-detection region of the background information as the extraction region, and the temperature of the non-detection region by the first background update process. It is more desirable to switch according to the variation degree of variation.

  According to the present invention, for a detection region, the update unit performs a second background update process for updating background information using a correction value obtained from the amount of change in temperature of the extraction region that is at least part of the non-detection region. Therefore, it becomes difficult to cause erroneous detection by responding to changes in the environmental temperature.

1 is a schematic diagram illustrating a configuration of an object detection device according to a first embodiment. 3 is a flowchart illustrating an operation of the object detection apparatus according to the first embodiment. FIG. 6 is an explanatory diagram illustrating an operation of the object detection device according to the first embodiment. FIG. 6 is an explanatory diagram illustrating an operation of the object detection device according to the first embodiment. FIG. 6 is an explanatory diagram illustrating an operation of the object detection device according to the first embodiment. FIG. 10 is an explanatory diagram illustrating an operation of an object detection device according to another configuration of the first embodiment. FIG. 10 is an explanatory diagram illustrating an operation of an object detection device according to another configuration of the first embodiment. FIG. 10 is an explanatory diagram illustrating an operation of an object detection device according to another configuration of the first embodiment. FIG. 10 is an explanatory diagram illustrating an operation of the object detection device according to the second embodiment.

  As shown in FIG. 1, the object detection apparatus 1 according to the present embodiment performs an operation such as a background difference, an acquisition unit 11 that acquires information from the sensor device 2, a storage unit 12 that stores information, and the object 3. A detection unit 13 that detects presence / absence and an output unit 14 that outputs a detection result are provided. Here, the object 3 is the object 3, but the object detection apparatus 1 of the present embodiment can detect any object as long as it is a heat source such as a traveling car.

  The sensor device 2 is formed by two-dimensionally arranging sensor elements (not shown) having sensitivity to infrared rays, such as a thermopile element, in a matrix, and information representing a temperature distribution in the detection range (sensing range) is two-dimensionally displayed. Output continuously as a thermal image (temperature distribution image). Each sensor element corresponds to each pixel of the thermal image, and hereinafter, the Celsius temperature of the area detected by each sensor element will be described as the pixel value of each corresponding pixel in the thermal image. In the example of FIG. 1, the object detection device 1 is separate from the sensor device 2. However, the present invention is not limited to this configuration, and the object detection device 1 may be integrally provided with the sensor device 2.

  The object detection apparatus 1 according to the present embodiment includes a computer. By causing a computer to execute a predetermined program, each function as an acquisition unit 11, a detection unit 13, and an update unit 15 described later is realized by the computer. The output unit 14 includes a device having a function of presenting the detection result of the detection unit 13 to the user, such as a monitor device or a printer device.

  The object detection device 1 acquires acquisition information representing the temperature distribution within the detection range from the sensor device 2 by the acquisition unit 11, and changes the acquisition information with respect to the background information stored in the storage unit 12 by the detection unit 13. Based on this, the presence or absence of the object 3 in the detection range is detected. In other words, the object detection device 1 records in the storage unit 12 in advance a thermal image representing the temperature distribution of the detection range when the object 3 does not exist in the detection range as a background image (background information). The object detection device 1 periodically acquires an acquired image (acquisition information) that is a thermal image representing the temperature distribution of the current detection range from the sensor device 2, and detects a difference between the acquired image and the background image. By the method, the detection unit 13 detects the object 3.

  However, if the background image acquired once is used as it is, the object detection device 1 will detect the temperature within the detection range that does not depend on the object 3 (hereinafter referred to as “environment temperature”) due to a change in time zone, air conditioning, or the like. There is a possibility that false detection will occur when changes occur.

  Therefore, the object detection apparatus 1 of the present embodiment includes an update unit 15 that repeatedly updates the background image stored in the storage unit 12. Here, the update unit 15 updates the background image every time the detection unit 13 performs the detection processing of the target object 3 by comparing the acquired image and the background image. The background image may be updated at a cycle of once every detection process. Further, the update unit 15 may update the background image at, for example, a predetermined time.

  Hereinafter, the operation of the object detection apparatus 1 will be described with reference to the flowchart of FIG. 2, focusing on the detection process of the object 3 by the detection unit 13 and the update process of the background image by the update unit 15. Moreover, FIGS. 3-5 has illustrated the process of the object detection apparatus 1 by taking the thermal image of 5 * 5 pixels as an example. In the following, the pixel value of each pixel in the image is specified by the coordinate position (x, y) of each pixel in the image and time t.

  First, the object detection device 1 acquires a thermal image output from the sensor device 2 in a state where the target object 3 is not present, by the acquisition unit 11, and sets each pixel value of the thermal image as background data Tb (x, y , T) is stored in advance in the storage unit 12 (S1). The background data Tb (x, y, t) is a pixel value of each pixel constituting the background image Tb.

  The object detection device 1 acquires the current temperature distribution data T (x, y, t) from the sensor device 2 by the acquisition unit 11 and stores the temperature distribution data T (x, y, t) in the storage unit 12. (S2). The temperature distribution data T (x, y, t) is a pixel value of each pixel constituting the acquired image T.

  As illustrated in FIG. 3, the detection unit 13 calculates the difference between the background data Tb (x, y, t) stored in the storage unit 12 and the temperature distribution data T (x, y, t) as difference data. Calculated as Td (x, y, t) (S3). That is, the difference data Td (x, y, t) is expressed by Equation 1. The difference data Td (x, y, t) is a pixel value of each pixel constituting the difference image Td. The detection unit 13 performs the calculation expressed by Equation 1 for all pixels of the acquired image T for each pixel.

それから、検出部１３は、差分データＴｄ（ｘ，ｙ，ｔ）を所定の閾値によって２値化する（Ｓ４）。このとき、検出部１３は、図３に例示するように、差分値が閾値以上の画素を対象物３が検出された「検出画素」とし（Ｓ５）、差分値が閾値より小さな画素を対象物３が検出されなかった「不検出画素」とする（Ｓ６）ように２値化し、検出画像Ｉを生成する。図３〜５の例では、斜線部が検出画素の集合からなる領域、つまり対象物３が検出された領域である検出領域Ｄ１を表し、それ以外の領域が不検出領域の集合からなる領域、つまり対象物３が検出されなかった領域である不検出領域Ｄ０を表している。検出画像Ｉを構成する各画素の画素値Ｉ（ｘ，ｙ，ｔ）は閾値Ｃを用いて数２にて表される。

Then, the detection unit 13 binarizes the difference data Td (x, y, t) with a predetermined threshold (S4). At this time, as illustrated in FIG. 3, the detection unit 13 sets a pixel having a difference value equal to or larger than a threshold as a “detection pixel” in which the target 3 is detected (S5), and sets a pixel having a difference value smaller than the threshold as the target. The detected image I is generated by binarizing so that “3” is detected as “undetected pixel” (S6). In the example of FIGS. 3 to 5, the shaded area represents a region composed of a set of detection pixels, that is, a detection region D1 in which the object 3 is detected, and the other region is a region composed of a set of non-detection regions. That is, it represents a non-detection area D0 that is an area where the object 3 has not been detected. A pixel value I (x, y, t) of each pixel constituting the detection image I is expressed by Equation 2 using a threshold value C.

つまり、Ｉ（ｘ，ｙ，ｔ）＝１となる画素（ｘ，ｙ）が検出画素であり、Ｉ（ｘ，ｙ，ｔ）＝０となる画素（ｘ，ｙ）が不検出画素である。検出画像Ｉは、検出部１３の検出結果として出力部１４から出力される。

That is, a pixel (x, y) where I (x, y, t) = 1 is a detection pixel, and a pixel (x, y) where I (x, y, t) = 0 is a non-detection pixel. . The detection image I is output from the output unit 14 as a detection result of the detection unit 13.

  Next, the object detection apparatus 1 performs, in the update unit 15, first background update processing for updating the background data Tb (x, y, t) for the non-detection region D0 including a set of non-detection pixels (S7). ). The update unit 15 corrects the background data Tb (x, y, t) stored in the storage unit 12 based on the latest temperature distribution data T (x, y, t) for all pixels in the non-detection region D0. Thus, the first background update process is performed. Specifically, as illustrated in FIG. 4, the update unit 15 includes the background data Tb (x, y, t) stored in the storage unit 12 and the current temperature distribution data T (x, y, t). The first background update process is performed by calculating a weighted average (weighted average) value. That is, the new background data Tb (x, y, t + 1) in the non-detection area D0 is expressed by Equation 3 using the weighted average parameter α. In the example of FIG. 4, the parameter α = 0.1. The update unit 15 performs the calculation expressed by Equation 3 for all the pixels in the non-detection region D0 for each pixel.

更新部１５は、第１の背景更新処理で得られた不検出領域Ｄ０の新たな背景データＴｂ（ｘ，ｙ，ｔ＋１）を用いて、後述の第２の背景更新処理で用いられる補正値を求める。補正値は、背景画像Ｔｂのうち不検出領域Ｄ０の少なくとも一部からなる抽出領域において、第１の背景更新処理で更新された温度の変化量から求められる。本実施形態では、不検出領域Ｄ０の全域を抽出領域とする。ここでは、更新部１５は、まず数４で表される更新値ΔＴｂ（ｘ，ｙ，ｔ）を、図５（ａ）に例示するように、画素ごとに不検出領域Ｄ０の全画素について求める（Ｓ８）。更新値ΔＴｂ（ｘ，ｙ，ｔ）は、背景画像Ｔｂにおいて第１の背景更新処理によって更新された画素値の変化量を表す値である。

The update unit 15 uses the new background data Tb (x, y, t + 1) of the non-detection area D0 obtained in the first background update process, and calculates a correction value used in the second background update process described later. Ask. The correction value is obtained from the amount of change in temperature updated in the first background update process in the extraction region consisting of at least a part of the non-detection region D0 in the background image Tb. In the present embodiment, the entire non-detection area D0 is set as the extraction area. Here, the update unit 15 first obtains the update value ΔTb (x, y, t) expressed by Equation 4 for all the pixels in the non-detection region D0 for each pixel as illustrated in FIG. 5A. (S8). The update value ΔTb (x, y, t) is a value representing the amount of change in the pixel value updated by the first background update process in the background image Tb.

続いて、更新部１５は、抽出領域である不検出領域Ｄ０の全画素における更新値ΔＴｂ（ｘ，ｙ，ｔ）の代表値を、補正値として求める（Ｓ９）。本実施形態では、補正値となる更新値ΔＴｂ（ｘ，ｙ，ｔ）の代表値は数５にて表される平均値ΔＴａｖとする。ただし、代表値は平均値に限らず、中央値や最頻値などであってもよい。なお、数５では、不検出領域Ｄ０に属する要素数（画素数）を｜Ｄ０｜として表している。

Subsequently, the update unit 15 obtains a representative value of the update value ΔTb (x, y, t) in all the pixels of the non-detection area D0 that is the extraction area as a correction value (S9). In the present embodiment, the representative value of the update value ΔTb (x, y, t) serving as the correction value is the average value ΔTav expressed by Equation 5. However, the representative value is not limited to the average value, and may be a median value or a mode value. In Equation 5, the number of elements (number of pixels) belonging to the non-detection region D0 is represented as | D0 |.

図５の例では、１９画素からなる不検出領域Ｄ０中に、更新値ΔＴｂ（ｘ，ｙ，ｔ）＝０．１の画素が１２画素存在するので、平均値ΔＴａｖは１．２／１９≒０．０６となる。

In the example of FIG. 5, since there are 12 pixels of the update value ΔTb (x, y, t) = 0.1 in the non-detection area D0 consisting of 19 pixels, the average value ΔTav is 1.2 / 19≈. 0.06.

  Then, the object detection apparatus 1 performs the second background update process in which the update unit 15 updates the background data Tb (x, y, t) for the detection region D1 including the set of detection pixels (S10). The update unit 15 uses the correction values (average value ΔTav) obtained in the processes S8 and S9 for all the pixels in the detection area D1, and uses the background data Tb (x, y, t) stored in the storage unit 12. By correcting, the second background update process is performed. Specifically, as illustrated in FIG. 5B, the update unit 15 adds the correction value to the current background data Tb (x, y, t) stored in the storage unit 12. The second background update process is performed. That is, the new background data Tb (x, y, t + 1) in the detection area D1 is expressed by Equation 6. The updating unit 15 performs the calculation represented by Equation 6 for all the pixels in the detection region D1 for each pixel.

要するに、更新部１５は、背景データＴｂ（ｘ，ｙ，ｔ）のうち、不検出領域Ｄ０については数３による第１の背景更新処理を行い、検出領域Ｄ１については数６による第２の背景更新処理を行う。更新部１５は、このようにして得られる新たな背景データＴｂ（ｘ，ｙ，ｔ＋１）によって、記憶部１２内の背景画像Ｔｂを随時書き換える。その結果、記憶部１２には、全画素（不検出領域Ｄ０および検出領域Ｄ１の両方の全画素）について値が更新された新たな背景データＴｂ（ｘ，ｙ，ｔ＋１）からなる背景画像Ｔｂ（ｔ＋１）が、背景画像Ｔｂとして記憶されることになる。なお、センサ装置２の検知範囲内に対象物３が存在しなければ、差分画像Ｔｄは検出領域Ｄ１を含まず、全域が不検出領域Ｄ０になるので、第１の背景更新処理のみで背景画像Ｔｂの全域が更新される。

In short, in the background data Tb (x, y, t), the update unit 15 performs the first background update process according to Equation 3 for the non-detection region D0, and the second background according to Equation 6 for the detection region D1. Perform update processing. The update unit 15 rewrites the background image Tb in the storage unit 12 as needed with the new background data Tb (x, y, t + 1) obtained in this way. As a result, the storage unit 12 stores a background image Tb (new background data Tb (x, y, t + 1)) with updated values for all pixels (all pixels in both the non-detection region D0 and the detection region D1). t + 1) is stored as the background image Tb. If the object 3 does not exist within the detection range of the sensor device 2, the difference image Td does not include the detection region D1, and the entire region becomes the non-detection region D0. Therefore, the background image is obtained only by the first background update process. The entire area of Tb is updated.

  Thereafter, the object detection apparatus 1 returns to the process of S2 and repeatedly executes the processes of S2 to S10 using the updated background image Tb.

  According to the object detection device 1 of the present embodiment described above, even when the environmental temperature changes while the target object 3 continues to exist at the same location, the update unit 15 detects that the target object 3 has been detected. Regarding the region D1, the background image Tb can be updated following the change in the environmental temperature. Therefore, even when the object 3 continues to exist in the same place for a long time, the background image Tb is correctly updated in response to a change in the environmental temperature, and the object detection device 1 uses the updated background image Tb. Subsequently, the detection process of the object 3 can be correctly performed.

  In short, the update unit 15 updates the background image Tb not only in the non-detection area D0 where the object 3 is not detected but also in the detection area D1 where the object 3 is detected. Can respond to changes in environmental temperature. For example, even when the ambient temperature is lowered while a person is still present at the same position, the background image Tb in the area where the person is present is also updated according to the change in the ambient temperature, so the object detection device 1 does not have a person. It is possible to avoid erroneous detection (missing). In addition, when the environmental temperature rises while a person is still present at the same position, the object detection device 1 may erroneously detect that there is a person in the area when the person leaves the area. Can be avoided.

  As a result, the object detection device 1 according to the present embodiment has an advantage that it is less likely to cause false detection by responding to a change in the environmental temperature, compared to a configuration in which the detection region D1 is not updated in the background image Tb. .

  By the way, in the above embodiment, the detection area D1 is a set of pixels (detection pixels) whose difference data is equal to or greater than a threshold value. However, the present invention is not limited to this example, and the detection area D1 is desirably enlarged appropriately. . That is, the detection area D1 only needs to include at least the entire area where the object 3 is detected by the detection unit 13, and is not limited to the area where the object 3 is detected. Similarly, the non-detection area D0 is an area other than the detection area D1, and is not limited to the area itself in which the object 3 is not detected. By enlarging the detection region D1, the following problems can be solved.

  That is, when the pixel resolution of the sensor device 2 is rough, an average low pixel value (temperature) is output for a pixel in which a portion where the object 3 exists and a portion where the object 3 does not exist are mixed in the same pixel. The difference data may be included in the non-detection area D0 without exceeding the threshold value. In this pixel, since the object 3 actually exists, the pixel value is higher than the original environmental temperature although it is less than the threshold value. Therefore, according to the object detection apparatus 1 of the above-described embodiment, the background data is updated to a temperature higher than the original environmental temperature by the update unit 15, which causes erroneous detection.

  On the other hand, when the detection area D1 is enlarged, a pixel in which a part where the object 3 exists and a part where the object 3 does not exist are included in the detection area D1. It is possible to avoid the background data being updated with the pixel value. Therefore, it is possible to avoid the background data being updated to a temperature higher than the original environmental temperature by the updating unit 15.

  Specifically, the enlargement process of the detection area D1 is performed on the pixel having the pixel value I (x, y, t) = 1 in the detection image I generated by the processes S4 to S6 ( Dilation). As a result, the detection area D1 of the difference image Td is expanded not only to a set of pixels (detection pixels) whose difference data Td (x, y, t) is equal to or greater than the threshold value as illustrated in FIG. The The dilation processing here is a general method as an image processing technique, and is a processing for enlarging the boundary line of pixel blocks having the same value and connected to each other by one pixel. By performing this expansion process, it is possible to include in the detection region D1 a pixel in which a portion where the object 3 exists and a portion where the target 3 does not exist in the same pixel.

  After performing the enlargement process of the detection area D1, the object detection apparatus 1 sets the area other than the detection area D1 as the non-detection area D0 and performs the first background update process and the second process as illustrated in FIGS. The background update process is executed (the above processes S7 to S10). 6-8, the shaded area represents the detection area D1, and the other areas represent the non-detection area D0.

(Embodiment 2)
The object detection device 1 according to the present embodiment is different from the object detection device 1 according to the first embodiment in the setting of the extraction region used when obtaining the correction value for the second background update process. In the following, the same components as those in the first embodiment are denoted by the same reference numerals, and the description thereof is omitted as appropriate.

  In the object detection device 1 according to the first embodiment, the entire non-detection region D0 is an extraction region, and the update unit 15 sets the representative values of the update values ΔTb (x, y, t) of all the pixels in the non-detection region D0. The correction value for the background update process 2 is used. Therefore, when the detection range of the sensor device 2 is wide and the change in environmental temperature is biased within the detection range, the following problems may occur. For example, when a heat source other than the object 3 such as a stove exists at a position away from the object (person) 3, the temperature around the heat source that is not related to the original environmental temperature of the detection region D1 is the second background. It may affect the correction value for the update process. Thereby, the background data of the detection area D1 may be updated to a temperature higher than the original environmental temperature.

  On the other hand, in this embodiment, the update unit 15 sets only a part of the non-detection area D0 in contact with the detection area D1 as an extraction area. That is, the update unit 15 calculates a representative value (for example, average value ΔTav) of the update value ΔTb (x, y, t) in the neighboring pixels adjacent to the detection region D1 in the non-detection region D0 as a correction value. A second background update process is performed using the value. Specifically, the update unit 15 performs expansion processing (n = 1, 2,...) On the detection image I n times, and the difference between the expanded region and the detection region D1 before the expansion. A representative value (correction value) of the update value ΔTb (x, y, t) is obtained by setting the region as an extraction region and targeting only pixels in the extraction region. Here, n is a parameter that determines the width (width) of the extraction region, and may be set according to the detection range and pixel resolution of the sensor device 2.

  According to the object detection device 1 of the present embodiment described above, since only a part of the non-detection area D0 that is in contact with the detection area D1 is an extraction area, it is biased toward changes in environmental temperature within the detection range. Even if there is, an appropriate value is calculated as the correction value for the second background update process. That is, the background data of the detection region D1 in which the object 3 is detected by the second background update process using the correction value is updated using the change in the environmental temperature in the vicinity of the object 3. It will be corrected to a new value. Therefore, even when the change in the environmental temperature is uneven within the detection range, it is possible to avoid the background data in the detection area D1 being updated to a temperature higher than the original environmental temperature in the second background update process. This configuration is particularly useful when the environmental temperature changes tend to be biased within the detection range, such as when the detection range of the sensor device 2 is wide.

  Further, as illustrated in FIG. 9, when there are a plurality of detection pixel clusters connected to each other, the updating unit 15 detects, for each cluster, a detection image Im (m = 1, 2,. ..) May be temporarily generated, and the extraction region may be set individually for each detected image Im. That is, the update unit 15 generates detection images Im as many as the number of detection pixel clusters, individually performs expansion processing for each detection image Im n times (twice in FIG. 9), and expands the region after expansion and before expansion. The difference areas from the detection areas D11 and D12 are the extraction areas D13 and D14. In FIG. 9, the detection areas D11 and D12 that are detection pixel clusters and the extraction areas D13 and D14 that are correction value calculation targets are distinguished by the direction of diagonal lines.

  In this case, the update unit 15 calculates a correction value for each of the extraction regions D13 and D14, and updates the background data of each detection region D11 and D12 using each correction value. That is, in the example of FIG. 9, the correction value calculated from the extraction region D13 of the detection image I1 is used for background data update of the detection region D11 of the detection image I1, and is calculated from the extraction region D14 of the detection image I2. The value is used to update the background data of the detection area D12 of the detection image I2.

  In the present embodiment, the expansion processing in the image processing technique is simply adopted as a method of determining the extraction region. However, depending on the arrangement of the sensor device 2, the pixel adjacent to the detection pixel is in the vicinity of the object 3 in the real space. It may not be. When it is known in advance that the arrangement of the sensor device 2 is such an arrangement, for example, when the sensor device 2 is arranged obliquely downward, the update unit 15 includes the object 3 among the detection pixels. Only the peripheral pixels of the foot area that becomes the contact point between the floor and the floor may be used as the extraction area.

  Other configurations and functions are the same as those of the first embodiment.

(Embodiment 3)
The object detection device 1 of the present embodiment is different from the object detection device 1 of the second embodiment in the setting of the extraction region used when obtaining the correction value for the second background update process. In the following, the same components as those in the second embodiment are denoted by the same reference numerals, and description thereof is omitted as appropriate.

  In the present embodiment, the updating unit 15 performs the first mode in which only a part of the non-detection area D0 that is in contact with the detection area D1 is the extraction area, and the first mode in which the entire non-detection area D0 is the extraction area. The two operation modes, i.e., the two modes, can be switched. The update unit 15 switches between the first mode and the second mode according to the variation degree of the temperature change amount of the non-detection region D0 by the first background update process.

  Specifically, the update unit 15 determines whether or not the value of the update value ΔTb (x, y, t) calculated with the first background update process is substantially uniform for all the pixels in the non-detection region D0. If it is determined that it is not uniform, it operates in the first mode, and if it is determined that it is not uniform, it operates in the second mode. Here, for example, the update unit 15 compares the difference value between the maximum value and the minimum value of the update value ΔTb (x, y, t) in the non-detection region D0 with a predetermined threshold value, and if the difference value is smaller than the threshold value, If it is uniform and is equal to or greater than the threshold value, it is determined that it is not uniform. For example, the update unit 15 obtains a standard deviation of the update value ΔTb (x, y, t) and determines whether or not the value is substantially uniform depending on whether the value is within a predetermined threshold range. May be.

  According to the object detection device 1 of the present embodiment described above, since the extraction region that is a correction value calculation target changes according to the variation degree of the temperature change amount of the non-detection region D0, the update unit 15 uses the optimum method. The correction value can be calculated with.

  In other words, the first mode in which a part of the area in contact with the detection area D1 is an extraction area is effective when the environmental temperature is uneven in the detection range, but the number of pixels to be calculated is reduced. There is a possibility that the correction value varies due to the increase in calculation load due to the expansion process. On the other hand, the second mode in which the entire non-detection area D0 is the extraction area is not suitable when there is a bias in the change in the environmental temperature within the detection range, but the variation in correction value due to a decrease in the number of pixels to be calculated. In addition, there is an advantage that the calculation load associated with the expansion process does not increase. Since the update unit 15 of the present embodiment switches between the first mode and the second mode according to the variation degree of the temperature change amount in the non-detection region D0, the correction value can always be calculated by an optimum method.

  Note that the area that the update unit 14 uses as the extraction area in the first mode is not limited to a part of the non-detection area D0 that is in contact with the detection area D1, but is a part of the non-detection area D0. Good.

  Other configurations and functions are the same as those of the second embodiment.

DESCRIPTION OF SYMBOLS 1 Object detection apparatus 2 Sensor apparatus 3 Object 11 Acquisition part 12 Storage part 13 Detection part 15 Update part

Claims (3)

An acquisition unit for acquiring acquisition information representing a temperature distribution in the detection range, a storage unit for storing background information indicating a temperature distribution in the detection range when no object is present in the detection range, and the background information A detection unit that detects the presence or absence of the object in the detection range based on a change in the acquired information, and an update unit that repeatedly updates the background information stored in the storage unit,
The update unit divides the background information into a detection region including a region where the object is detected by the detection unit, and a non-detection region including a region other than the detection region,
For the non-detection area, a first background update process is performed to update the background information based on the acquired information. An object detection apparatus that executes a second background update process for updating the background information using a correction value obtained from the amount of change in temperature by one background update process.   The update unit obtains the correction value from the amount of change in temperature of the extraction region, with a part of the non-detection region in contact with the detection region as the extraction region. The object detection apparatus described in 1. The update unit uses a partial area of the non-detection area of the background information as the extraction area, a first mode for obtaining the correction value from the amount of change in temperature of the extraction area, and the non-detection of the background information The second mode for obtaining the correction value from the amount of change in the temperature of the extraction region with the entire region as the extraction region is set to the degree of variation in the amount of change in the temperature of the non-detection region by the first background update process. The object detection apparatus according to claim 1, wherein the object detection apparatus is switched according to the selection.

Images